Walker A Peer Tutoring Addition

From LearnLab
Revision as of 16:26, 15 February 2007 by Vanlehn (Talk | contribs) (Abstract)

Jump to: navigation, search

Collaborative Extensions to the Cognitive Tutor Algebra: A Peer Tutoring Addition

Erin Walker, Bruce McLaren, Ken Koedinger, and Nikol Rummel


We wish to evaluate the potential benefits of combining peer tutoring with a cognitive tutor. The study compares an experimental condition to control conditions. In the experimental condition, students first prepare to tutor, and then take turns tutoring each other using the cognitive tutor interface. They receive help from the cognitive tutor as they collaborate. In the first control, students prepare to tutor and tutor each other, but receive no help from the cognitive tutor. In the second control, students solve problems alone with the help of the cognitive tutor.

In this in vivo experiment in the Algebra LearnLab, three classes will be used with roughly 20 students per class. The study will last a week, and students will work on problems from an equation solving unit. Robust learning will be measured.

Background and Significance


  • Cognitive tutor – the computer tutor agent
  • Peer tutor – the human tutor agent
  • Student – the person being tutored

Research question

Do instructional activities that involve three agents (a student, a peer tutor, and a computer tutor) increase robust learning compared to instructional activities that involve two agents (student and peer tutor, student and computer tutor)?

Independent variables

  • Number of agents (2 or 3)
  • Additional agent(s) type (peer tutor, computer tutor, or both)

The condition where a student interacts with a peer tutor and computer tutor is considered the “peer + cognitive tutoring” condition. The condition where a student interacts only with a peer tutor is considered the “peer tutoring” condition. The condition where a student interacts only with a computer tutor is considered the “cognitive tutoring” condition.


When students attempt to solve a problem with the help of both a peer tutor and cognitive tutor, they should show more robust learning than when they problem solve with the help of only one of the tutors. A student/cognitive tutor collaboration should produce similar levels of robust learning as a student/peer tutor collaboration, as long as the peer tutor is sufficiently prepared.

Dependent variables

  • Near transfer, immediate: During training, student progress on training problems will be analyzed.
  • Near transfer, retention: Students will be given a posttest immediately after the study on similar problems
  • Acceleration of future learning: Student learning on future equation solving units will be measured.


This study is part of the Interactive Communication cluster, and its hypothesis is a specialization of the IC cluster’s central hypothesis. We assume that the instruction is in or above the student’s ZPD. Here is what we expect to happen in the three conditions:

  • In the cognitive tutoring condition, the cognitive tutor’s help should increase learning compared to student learning alone. However, the student does not have the benefit of natural language interaction with another agent, and learning will not be optimized.
  • In the peer tutoring condition, the peer tutor’s help should increase learning compared to a student learning alone. However, the student may make too many errors and/or require too much communication with the second agent. The peer tutor may not be able to provide the student with the help he/she needs. Learning will not be optimized.
  • In the cognitive+peer tutoring condition, the peer tutor’s help should provide the additional benefits of interactive communication, and the cognitive tutor’s help should ensure that both the student and the peer tutor get the problem solving assistance they need. The support of the cognitive tutor and interaction with the student will also increase the peer tutor’s learning.

The target path will have the students learning-by-doing and learning-by-teaching, with the help of the computer agent so students do not make too many errors or require too much communication.


A previous study involving only peer tutoring suggested that cognitive tutoring is indeed necessary so that problems are in the ZPD of the student and peer tutor. The current study will be completed in the fall.

Annotated bibliography

Walker, E., Rummel, N., McLaren, B. M. & Koedinger, K. R. (submitted). The Student Becomes the Master: Integrating Peer Tutoring with Cognitive Tutoring. Submitted to the conference on Computer Supported Collaborative Learning (CSCL-07). Rutgers University, July 16-21, 2007.

Walker, E., Koedinger, K., McLaren, B. M., & Rummel, N. (2006). Cognitive tutors as research platforms: Extending an established tutoring system for collaborative and metacognitive experimentation. Lecture Notes in Computer Science, Volume 4053/2006. Proceedings of the 8th International Conference on Intelligent Tutoring Systems (pp. 207-216). Berlin: Springer

Walker, E. (2005). Mutual peer tutoring: A collaborative addition to the Algebra-1 Cognitive Tutor. Paper presented at the 12th International Conference on Artificial Intelligence and Education (AIED-05, Young Researchers Track), July, 2005, Amsterdam, the Netherlands.